Supplementary data 1
Reproducible report
Lucas Bravo1, Milena Cano1, Mauricio F. Landaeta2,3, Sergio Navarrete4,5, Simone Baldanzi1,2
1 Laboratorio de Ecofisiología y Ecología Evolutiva Marina (e°CO2lab), Facultad de Ciencias del mar y de Recursos Naturales, Universidad de Valparaíso, Viña del Mar, Chile.
2 Centro de Observación Marino para Estudios de Riesgos del Ambiente Costero (COSTA-R), Universidad de Valparaíso, Chile
3 Laboratorio de Ictiología e Interacciones Biofíscas (LABITI), Facultad de Ciencias, Universidad de Valparaíso, Valparaiso, Chile
4 Estación Costera de Investigaciones Marinas (ECIM), Pontificia Universidad Católica de Chile, Las Cruces, Chile.
5 Center for Applied Ecology and Sustainability (CAPES), COPAS COASTAL, Instituto Milenio en Socioecología Costera, SECOS, Pontificia Universidad Católica de Chile, Santiago, Chile.
This document reproduces and supports all data work and statistical analysis of the larval survival, geometric morphometrics and swimming activity in the paper “The effect of rearing temperatures on the swimming performance, survival and morphometrics changes of zoea 1 Taliepus dentatus”.
Libraries
#Mortality libraries
library(cowplot)
library(scales)
library(dplyr)
library(gridExtra)
library(tidyverse)
library(ggsci)
library(Rmisc)
library(survminer)
library(survival)
library(MASS)
library(here)
#Morphometrics Libraries
library(Morpho)
library(geomorph)
library(shapes)
library(StereoMorph)
library(readr)
library(ggplot2)
#Swimming libraries
library(tidyverse)
library(lme4)
library(plyr)
library(lmerTest)
library(lme4)
library(car)
library(sjPlot)
library(effects)
library(cowplot)
library(emmeans)
library(sjPlot)
library(multcomp)
library(rlist)Timeseries of environmental temperature
montemar<-read.table(here("Montemar-minidot", "miniDOT.dat"))
#MAKING DATETIME COLUMN and ADDING COLUMNS GROUPS
montemar_date <- montemar %>%
unite("date", V1:V3, sep = "-")
montemar_date$date<-anytime::anydate(montemar_date$date)
montemar_date$jd <- yday(montemar_date$date)
montemar_date$day<-day(montemar_date$date)
montemar_date$month<-month(montemar_date$date)
montemar_date$year<-year(montemar_date$date)
#mean temperature by days
Day_temp<-montemar_date %>%
group_by(year,month, day) %>%
dplyr::summarize(mean_temp = mean(V6), SD=sd(V6))
Day_temp<-Day_temp %>% mutate(Date = make_date(year, month, day))
Day_temp<-Day_temp[1:365,] #RUN TO TAKE ONLY FIRST YEAR (2021 JAN-2022 JAN)
Day_temp$month<-as.factor(Day_temp$month)
#mean temperature by month
Month_temp <- montemar_date %>%
group_by(year,month) %>%
dplyr::summarize(mean_temp = mean(V6, na.rm=T), SD=sd(V6,na.rm=T), n=n())
Month_temp## # A tibble: 22 × 5
## # Groups: year [2]
## year month mean_temp SD n
## <dbl> <dbl> <dbl> <dbl> <int>
## 1 2021 1 15.4 1.07 1342
## 2 2021 2 14.5 1.07 4032
## 3 2021 3 14.6 0.953 4464
## 4 2021 4 13.9 0.611 4320
## 5 2021 5 14.1 0.888 4464
## 6 2021 6 12.6 0.521 4320
## 7 2021 7 12.2 0.159 4464
## 8 2021 8 12.5 0.198 4464
## 9 2021 9 12.5 0.410 4320
## 10 2021 10 12.2 0.644 4464
## # ℹ 12 more rowsggplot(Day_temp, aes(x=Date, y=mean_temp)) + geom_point() + geom_line()+ theme_bw() + scale_x_date(date_labels = "%b", date_breaks = "1 month") +
geom_errorbar(aes(ymin=Day_temp$mean_temp-SD, ymax=Day_temp$mean_temp+SD),
width=.2, position=position_dodge(0.05),alpha=0.5) +
labs(y="Temperature (C°)")+ theme(plot.margin = margin(0.2,0.5,0.2,0.2, "cm")) + labs(x=element_blank())
Mortality
1. Import of data base and definition of factors
#data are available in the online version of the article, remember to import environment
Mort <- read.delim(here::here("mortality/mort.txt"))
Mort$bino<-Mort$alive/100
#Temperature as factor
Mort$Temperature<-as.factor(Mort$Temperature)2. Data summary
# SD, SE, CI and MEAN
LT50_summary<- summarySE(Mort, measurevar="bino",
groupvars=c("Temperature","Day"))
head(LT50_summary)## Temperature Day N bino sd se ci
## 1 12 1 3 1.0000000 0.000000000 0.000000000 0.00000000
## 2 12 2 3 0.9766667 0.005773503 0.003333333 0.01434218
## 3 12 3 3 0.5000000 0.115325626 0.066583281 0.28648474
## 4 12 4 3 0.3066667 0.210792157 0.121700908 0.52363675
## 5 12 5 3 0.2966667 0.202319879 0.116809436 0.50259044
## 6 12 6 3 0.2966667 0.202319879 0.116809436 0.502590443. Generalized Lineal Model (GLM) analysis and LT50%
#separation of the table by temp
mort_12<-LT50_summary[1:11, ]
mort_15<-LT50_summary[12:22, ]
mort_17<-LT50_summary[23:32,]
#GLM analysis for 12°C
model1<-glm(data=mort_12, bino ~ Day, family=quasibinomial())
summary(model1)##
## Call:
## glm(formula = bino ~ Day, family = quasibinomial(), data = mort_12)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.5179 -0.2999 -0.0738 0.3400 0.5964
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.1710 0.6380 3.403 0.00784 **
## Day -0.4791 0.1117 -4.289 0.00202 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 0.1436587)
##
## Null deviance: 5.4993 on 10 degrees of freedom
## Residual deviance: 1.5399 on 9 degrees of freedom
## AIC: NA
##
## Number of Fisher Scoring iterations: 5#GLM analysis for 15°C
model2<-glm(data=mort_15, bino ~ Day, family=quasibinomial())
summary(model2)##
## Call:
## glm(formula = bino ~ Day, family = quasibinomial(), data = mort_15)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.185582 -0.053923 -0.005199 0.088830 0.186640
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 5.84038 0.37027 15.77 7.28e-08 ***
## Day -0.89940 0.05508 -16.33 5.39e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 0.01151942)
##
## Null deviance: 8.21286 on 10 degrees of freedom
## Residual deviance: 0.13672 on 9 degrees of freedom
## AIC: NA
##
## Number of Fisher Scoring iterations: 6#GLM analysis for 17°C
model3<-glm(data=mort_17, bino ~ Day, family=quasibinomial())
summary(model3)##
## Call:
## glm(formula = bino ~ Day, family = quasibinomial(), data = mort_17)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.287545 -0.221565 -0.146333 -0.006668 0.263467
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 6.6331 0.9512 6.974 0.000116 ***
## Day -1.1205 0.1559 -7.186 9.37e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for quasibinomial family taken to be 0.05103717)
##
## Null deviance: 8.41356 on 9 degrees of freedom
## Residual deviance: 0.38809 on 8 degrees of freedom
## AIC: NA
##
## Number of Fisher Scoring iterations: 6#Dose.p function to find moment at which 50% of the population has died (LT50%) based in the glm models made above
#LT50% for 12°C
dose.p(model1, p = 0.5)## Dose SE
## p = 0.5: 4.531844 0.6121778#LT50% for 15°C
dose.p(model2, p = 0.5)## Dose SE
## p = 0.5: 6.493628 0.11405#LT50% for 17°C
dose.p(model3, p = 0.5)## Dose SE
## p = 0.5: 5.919705 0.2142654# LT50% values with their CI
lt50<-c(4.531844, 6.493628, 5.919705)
CI<-c(0.6121778*1.95, 0.11405*1.95, 0.2142654*1.95)
temp<-c(12, 15,17)
data_final<-data.frame(temp,lt50, CI)
data_final$temp<-as.factor(data_final$temp)
data_final## temp lt50 CI
## 1 12 4.531844 1.1937467
## 2 15 6.493628 0.2223975
## 3 17 5.919705 0.41781754. Kaplan-Meier Analysis
#Vector for creating correct data table
repV<- Mort$dead
repV[repV=="0"]<-1
#Datatable KAP for analysis
KAP<-Mort %>%
mutate(count = repV) %>%
uncount(count)
# Adding Status column (binary in which 1=no death; 2=death on that day)
KAP$STATUS <- ifelse (KAP$dead==0, 1, 2)
survObj <- Surv(time=KAP$Day, event=KAP$STATUS==2)
#Kaplan-Meier analysis
sfit <- survfit(Surv(Day, STATUS)~Temperature, data=KAP)
summary(sfit)## Call: survfit(formula = Surv(Day, STATUS) ~ Temperature, data = KAP)
##
## Temperature=12
## time n.risk n.event survival std.err lower 95% CI upper 95% CI
## 2 308 7 0.9773 0.00849 0.9608 0.9941
## 3 301 143 0.5130 0.02848 0.4601 0.5720
## 4 158 58 0.3247 0.02668 0.2764 0.3814
## 5 100 3 0.3149 0.02647 0.2671 0.3713
## 8 90 6 0.2939 0.02605 0.2471 0.3497
## 9 83 11 0.2550 0.02511 0.2102 0.3093
## 10 72 57 0.0531 0.01328 0.0325 0.0867
## 11 15 15 0.0000 NaN NA NA
##
## Temperature=15
## time n.risk n.event survival std.err lower 95% CI upper 95% CI
## 3 297 21 0.9293 0.0149 0.9006 0.9589
## 4 276 11 0.8923 0.0180 0.8577 0.9282
## 5 265 17 0.8350 0.0215 0.7939 0.8783
## 6 248 66 0.6128 0.0283 0.5598 0.6708
## 7 182 69 0.3805 0.0282 0.3291 0.4399
## 8 113 48 0.2189 0.0240 0.1765 0.2713
## 9 65 28 0.1246 0.0192 0.0922 0.1684
## 10 37 27 0.0337 0.0105 0.0183 0.0619
## 11 10 10 0.0000 NaN NA NA
##
## Temperature=17
## time n.risk n.event survival std.err lower 95% CI upper 95% CI
## 2 298 9 0.96980 0.00991 0.95056 0.9894
## 3 289 12 0.92953 0.01483 0.90092 0.9590
## 4 276 8 0.90259 0.01719 0.86952 0.9369
## 5 268 12 0.86217 0.01999 0.82387 0.9023
## 6 256 80 0.59274 0.02851 0.53942 0.6513
## 7 176 105 0.23912 0.02475 0.19521 0.2929
## 8 71 63 0.02694 0.00940 0.01360 0.0534
## 9 8 5 0.01010 0.00580 0.00328 0.0311
## 10 3 1 0.00674 0.00475 0.00169 0.0268#Ploting Kaplan-meier results
ggsurvplot(sfit, conf.int=TRUE, pval=TRUE, risk.table=FALSE,
legend.labs=c("12°C", "15°C","17°C"),
legend.title="Temperature",
palette=(c("blue", "orange", "red")),
risk.table.height=.15, ylab="Survival", xlab="Day")
Fig 1. Kaplan-meier survival curve (lines) for each temperature with their confidence interval (shaded color)
Geometric Morphometrics
##Digitizing
curve.v1<- c("organo_dorsal","esq_izq_caparaza")
curve.v2<- c("punta_espina_dorsal","punta_espina_dorsal")
curve.v3<- c("curvaD_espina_dorsal","curvaI_espina_dorsal")
curve<- matrix(c(curve.v3,curve.v2, curve.v1), ncol = 3)
# take-off "#" of the digitizeimages function in order to visualize each picture digitalization on rmarkdown.
#digitizeImages(image.file="morphometric larva/ZOEA1_FOTOS/12°C", landmarks.ref="morphometric larva/landmarks.txt", curves.ref = curve, shapes.file ="morphometric larva/zoea1")
#digitizeImages(image.file="morphometric larva/ZOEA1_FOTOS/15°C", landmarks.ref="morphometric larva/landmarks.txt", curves.ref = curve, shapes.file ="morphometric larva/zoea1")
#digitizeImages(image.file="morphometric larva/ZOEA1_FOTOS/17°C", landmarks.ref="morphometric larva/landmarks.txt", curves.ref = curve, shapes.file ="morphometric larva/zoea1")1. Import Digitized pictures, grouping files and links for landmarks
#data are available in the online version of the article, remember to import environment
DIGIDAT<-readShapes(here::here("morphometric/zoea1"))
# Telling R the database is from steremorph
array_data<-readland.shapes(DIGIDAT, nCurvePts= c(7,7))
# Importing traits
traitF<- read_delim("/Users/lucasb/Desktop/ECOLAB!/paper tesis/morphometric larva/trait1.txt",
delim = "\t", escape_double = FALSE,
trim_ws = TRUE)
head(traitF)## # A tibble: 6 × 5
## ID Hours Temp Replica Day
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 12_1_IND1_C 12 12 1 1
## 2 12_1_IND1_D 12 12 1 1
## 3 12_1_IND1_I 12 12 1 1
## 4 12_1_IND2_C 12 12 2 1
## 5 12_1_IND2_D 12 12 2 1
## 6 12_1_IND2_I 12 12 2 1# assigning columns as factors
traitF$Temp<- as.factor(traitF$Temp)
traitF$Replica<-as.factor(traitF$Replica)
traitF$Day<-as.factor(traitF$Day)
# importing file linksgeo to create lines connections between landmarks and semilandmarks
linksgeo <- read_delim("/Users/lucasb/Desktop/ECOLAB!/paper tesis/morphometric larva/links.txt", delim = "\t",
escape_double = FALSE, col_names = FALSE,
trim_ws = TRUE)
linksgeo<-as.matrix(linksgeo)2. Generalized Procrustes Analysis (GPA)
#Generalized Procrustes Analysis
GPA<- gpagen(array_data,ProcD=FALSE, print.progress = FALSE, curves=array_data$curves)
#procD=false, bending energy is the correct criterion for optimizing the positions of semilandmarks
gdf<- geomorph.data.frame(GPA, Temp=traitF$Temp, replica=traitF$Replica)3. Principal Components Analysis (PCA)
# Mean shape of all the individuals and adding links and numbers for each landmark and semilandmark
msh <- mshape(GPA$coords)
plot(msh, links = linksgeo)
# plot of all landmarks and semilandmarks digitized in grey and mean shape in black
plotAllSpecimens(GPA$coords, links = linksgeo, plot.param = list(txt.cex = 1.5, txt.col = "#D53E4F", pt.bg = "#BDBDBD", link.col = "black", mean.bg = "black"))
# Principal Components Analysis
PCA<-gm.prcomp(GPA$coords)
summary(PCA)##
## Ordination type: Principal Component Analysis
## Centering by OLS mean
## Orthogonal projection of OLS residuals
## Number of observations: 57
## Number of vectors 33
##
## Importance of Components:
## Comp1 Comp2 Comp3 Comp4
## Eigenvalues 0.001663143 0.0009952127 0.0007084004 0.0005026865
## Proportion of Variance 0.344256689 0.2060006993 0.1466329435 0.1040518979
## Cumulative Proportion 0.344256689 0.5502573882 0.6968903317 0.8009422296
## Comp5 Comp6 Comp7 Comp8
## Eigenvalues 0.0003263609 0.0002361839 0.0001091267 8.809108e-05
## Proportion of Variance 0.0675539785 0.0488880929 0.0225883065 1.823412e-02
## Cumulative Proportion 0.8684962081 0.9173843010 0.9399726075 9.582067e-01
## Comp9 Comp10 Comp11 Comp12
## Eigenvalues 0.0000538088 4.697673e-05 2.731572e-05 0.0000227029
## Proportion of Variance 0.0111379705 9.723791e-03 5.654126e-03 0.0046993105
## Cumulative Proportion 0.9693446944 9.790685e-01 9.847226e-01 0.9894219213
## Comp13 Comp14 Comp15 Comp16
## Eigenvalues 0.0000169546 8.878019e-06 8.338615e-06 4.809795e-06
## Proportion of Variance 0.0035094592 1.837676e-03 1.726023e-03 9.955873e-04
## Cumulative Proportion 0.9929313806 9.947691e-01 9.964951e-01 9.974907e-01
## Comp17 Comp18 Comp19 Comp20
## Eigenvalues 3.339880e-06 2.629547e-06 2.110745e-06 1.387480e-06
## Proportion of Variance 6.913272e-04 5.442942e-04 4.369066e-04 2.871968e-04
## Cumulative Proportion 9.981820e-01 9.987263e-01 9.991632e-01 9.994504e-01
## Comp21 Comp22 Comp23 Comp24
## Eigenvalues 1.221668e-06 7.846171e-07 5.547431e-07 4.480001e-08
## Proportion of Variance 2.528751e-04 1.624092e-04 1.148272e-04 9.273227e-06
## Cumulative Proportion 9.997033e-01 9.998657e-01 9.999805e-01 9.999898e-01
## Comp25 Comp26 Comp27 Comp28
## Eigenvalues 1.811106e-08 1.023950e-08 8.629042e-09 4.528716e-09
## Proportion of Variance 3.748837e-06 2.119490e-06 1.786139e-06 9.374063e-07
## Cumulative Proportion 9.999935e-01 9.999956e-01 9.999974e-01 9.999984e-01
## Comp29 Comp30 Comp31 Comp32
## Eigenvalues 3.781360e-09 2.209234e-09 1.190465e-09 7.020675e-10
## Proportion of Variance 7.827099e-07 4.572929e-07 2.464163e-07 1.453221e-07
## Cumulative Proportion 9.999992e-01 9.999996e-01 9.999999e-01 1.000000e+00
## Comp33
## Eigenvalues 9.145980e-19
## Proportion of Variance 1.893141e-16
## Cumulative Proportion 1.000000e+00# Principal components analysis by replica
AnError<-procD.lm(coords~replica, data = gdf, iter = 999, RRPP = TRUE)
summary(AnError)##
## Analysis of Variance, using Residual Randomization
## Permutation procedure: Randomization of null model residuals
## Number of permutations: 1000
## Estimation method: Ordinary Least Squares
## Sums of Squares and Cross-products: Type I
## Effect sizes (Z) based on F distributions
##
## Df SS MS Rsq F Z Pr(>F)
## replica 18 0.229134 0.0127297 0.84694 11.682 12.72 0.001 **
## Residuals 38 0.041408 0.0010897 0.15306
## Total 56 0.270542
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Call: procD.lm(f1 = coords ~ replica, iter = 999, RRPP = TRUE, data = gdf)(0.0127297/0.0138194)*100## [1] 92.11471# Error explained by miss-digitazion (%)
(0.0010897/0.0138194)*100 ## [1] 7.885292# Taking out PCA coordinates to create a plot with ggplot
df_out <- as.data.frame(PCA$x)
# ggplot for the PC1 and PC2
PCA1<-ggplot(data=df_out, aes(x=Comp1, y=Comp2, col=
gdf$Temp)) + geom_point(size=3, show.legend = FALSE) + theme_light() + labs(x="PC1=34.4%", y="PC2=20.6%") + scale_color_manual(values =c("blue","orange", "red"))
PCA1
Fig 3a. First (PC1) and second (PC2) principal components distribution with their explained variance (%)
# Plots of the mean shape (grey) against mínimum shape (black) of PC1 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp1$min, method = "points",
links = linksgeo)
# Plots of the mean shape (grey) against maximum shape (black) of PC1 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp1$max,method = "points",
links = linksgeo)
# Plots of the mean shape (grey) against mínimum shape (black) of PC2 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp2$min, method = "points",
links = linksgeo)
# Plots of the mean shape (grey) against maximum shape (black) of PC2 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp2$max,method = "points",
links = linksgeo)# ggplot for the PC1 and PC3
PCA2<-ggplot(data=df_out, aes(x=Comp1, y=Comp3, col=
gdf$Temp))+geom_point(size=3) + theme_light() +
labs(x="PC1=34.4%", y="PC3=14.6%", shape="Day", col="Temperature")+ scale_color_manual(values =c("blue", "orange", "red"))
PCA2
Fig 3b. First (PC1) and second (PC3) principal components distribution with their explained variance (%)
# Plots of the mean shape (grey) against mínimum shape (black) of PC1 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp1$min, method = "points",
links = linksgeo)
# Plots of the mean shape (grey) against maximum shape (black) of PC1 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp1$max,method = "points",
links = linksgeo)
# Plots of the mean shape (grey) against mínimum shape (black) of PC3 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp3$min, method = "points",
links = linksgeo)
# Plots of the mean shape (grey) against maximum shape (black) of PC3 (with default magnification=1)
plotRefToTarget(msh, PCA$shapes$shapes.comp3$max,method = "points",
links = linksgeo)4. Canonical Variance Analysis (CVA)
# Column to the Database to later use as grouping
traitF$grouppp<-paste(traitF$Temp, traitF$Day)
# Canonical variance analysis using bonferroni p-value adjustment
cva1<-CVA(GPA$coords, groups = traitF$Temp, p.adjust.method = "bonferroni",rounds = 1000, robust = "classical")## singular Covariance matrix: General inverse is used. Threshold for zero eigenvalue is 1e-10# P-values (euclides) obtained between groups
(cva1[["Dist"]][["probsEuclid"]])## 12 15
## 15 0.002997003
## 17 0.095904096 0.002997003# Procrustes distances (euclides) between groups
(cva1[["Dist"]][["GroupdistEuclid"]])## 12 15
## 15 0.06321077
## 17 0.03565455 0.04666779Cva_out<-as.data.frame(cva1$CVscores)
consensus<-as.data.frame(cva1$groupmeans)
Cva_out$temp<-gdf$Temp
ggplot(data=Cva_out, aes(x=`CV 1`, y=`CV 2`, col=temp)) + geom_point() + stat_ellipse(level = 0.95) + coord_fixed() +
labs(x="Canonical variable 1 (76.4%)", y= "Canonical variable 2 (23.6%)", col= "Temperature") + theme_classic() + scale_color_manual(values =c("blue", "orange", "red"), labels=c("12°C","15°C","17°C"))
plotRefToTarget(cva1[["Grandm"]], cva1[["groupmeans"]][,,1], method ="points", links = linksgeo, axes = TRUE, label = FALSE, gridPars = gridPar(tar.pt.bg = "blue", tar.link.col="blue",tar.link.lwd=2, link.col = "grey", pt.bg = "grey", link.lwd = 2, pt.size = 1, tar.pt.size = 1.5))
plotRefToTarget(cva1[["Grandm"]], cva1[["groupmeans"]][,,2], method ="points", links = linksgeo, axes = TRUE, label = FALSE, gridPars = gridPar(tar.pt.bg = "orange", tar.link.col="orange",tar.link.lwd=2, link.col = "grey", pt.bg = "grey", link.lwd = 2, pt.size = 1, tar.pt.size = 1.5))
plotRefToTarget(cva1[["Grandm"]], cva1[["groupmeans"]][,,3], method ="points", links = linksgeo, axes = TRUE, label = FALSE, gridPars = gridPar(tar.pt.bg = "red", tar.link.col="red",tar.link.lwd=2, link.col = "grey", pt.bg = "grey", link.lwd = 2, pt.size = 1, tar.pt.size = 1.5))
Swimming Performance
1. Import database and Tyding
#data are available in the online version of the article, remember to import environment
list_of_files <- list.files(path = here::here("swimming R/NEW TRIAL/data"), recursive = TRUE,pattern = "\\.txt$", full.names = TRUE)
#Lapply allows to import all the files in one list
loop_full<-lapply(list_of_files, read.delim2)
names(loop_full)<-list_of_files
loop_full<-lapply(loop_full, function(x) { x["frame"] <- NULL; x })mean_t1_12<-list()
# for loop to make same functions to all files of 12°C
for (i in 1:14){ a=loop_full[[i]][,]
a$x<-as.numeric(a$x)
a$y<-as.numeric(a$y)
df<-a %>%
group_by(X) %>%
mutate(XX = x - lag(x)) %>% #Distance difference in x axis between two points
mutate(Y = y - lag(y)) #Distance difference in y axis between two points
df$time<-as.POSIXct(df$time, format="%M:%OS")
df$tdif<-ifelse(difftime(df$time,lag(df$time), units = "secs")<1,difftime(df$time,lag(df$time), units = "secs"),NA)
z1<-aggregate(df$tdif, by=list(Trajectory=df$X), FUN=sum, na.rm=T)
df$iv <- (sqrt((df$XX)^2+(df$Y)^2))/0.125 #Instant velocity
z1$IV <- aggregate(df$iv, list(df$X), FUN = mean, na.rm=TRUE)[,2] #Mean velocity by larvae
mean_t1_12[[i]]<-z1
}
names(mean_t1_12)<-list_of_files[1:14] #rownames
# Tidying columns Temp, aquarium and larva.
mean_t1_12<- list.rbind(mean_t1_12)
mean_t1_12$larva<-c(1,2,2,3,3,3,3,4,4,4,5,5,5,5,1,1,2,2,3,3,3,4,4,4,5,1,2,2,2,4,5,5,5)
mean_t1_12$aquarium<-c("A","A", "A", "A", "A", "A" ,"A", "A", "A", "A", "A", "A", "A", "A", "B", "B", "B", "B","B", "B", "B", "B", "B", "B", "B", "C", "C", "C", "C", "C", "C", "C", "C")
mean_t1_12$temp<-rep(12)
mean_t1_12<-mean_t1_12[, c(6,5,4, 1, 2, 3)]
colnames(mean_t1_12)<-c("Temp","aquarium","Larva","Trajectory","Time","IV")mean_t1_15<-list()
# for loop to make same functions to all files of 15°C
for (i in 15:25){ a=loop_full[[i]][,]
a$x<-as.numeric(a$x)
a$y<-as.numeric(a$y)
df<-a %>%
group_by(X) %>%
mutate(XX = x - lag(x)) %>% #Distance difference in x axis between two points
mutate(Y = y - lag(y)) #Distance difference in y axis between two points
df$time<-as.POSIXct(df$time, format="%M:%OS")
df$tdif<-ifelse(difftime(df$time,lag(df$time), units = "secs")<1,difftime(df$time,lag(df$time), units = "secs"),NA)
z1<-aggregate(df$tdif, by=list(Trajectory=df$X), FUN=sum, na.rm=T)
df$iv <- (sqrt((df$XX)^2+(df$Y)^2))/0.125 #Instant velocity
z1$IV <- aggregate(df$iv, list(df$X), FUN = mean, na.rm=TRUE)[,2] #Mean velocity by larvae
mean_t1_15[[i]]<-z1
}
mean_t1_15<-mean_t1_15[c(15:25)]
names(mean_t1_15)<-list_of_files[15:25] #rownames
# Tidying columns Temp, aquarium and larva with rownames
mean_t1_15<- list.rbind(mean_t1_15)
mean_t1_15$larva<-c(1,1,1,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,5,5,1,1,1,1,1,2,2,2,2,2,2,2,3,1,2,2,2,2,2,3,3,5,5)
mean_t1_15$aquarium<-c("A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "C", "C", "C","C", "C", "C", "C", "C", "C", "C")
mean_t1_15$temp<-rep(15)
mean_t1_15<-mean_t1_15[, c(6,5,4, 1, 2, 3)]
colnames(mean_t1_15)<-c("Temp","aquarium","Larva","Trajectory","Time","IV")mean_t1_17<-list()
# for loop to make same functions to all files of 17°C
for (i in 26:39){ a=loop_full[[i]][,]
a$x<-as.numeric(a$x)
a$y<-as.numeric(a$y)
df<-a %>%
group_by(X) %>%
mutate(XX = x - lag(x)) %>% #Distance difference in x axis between two points
mutate(Y = y - lag(y)) #Distance difference in y axis between two points
df$time<-as.POSIXct(df$time, format="%M:%OS")
df$tdif<-ifelse(difftime(df$time,lag(df$time), units = "secs")<1,difftime(df$time,lag(df$time), units = "secs"),NA)
z1<-aggregate(df$tdif, by=list(Trajectory=df$X), FUN=sum, na.rm=T)
df$iv <- (sqrt((df$XX)^2+(df$Y)^2))/0.125 #Instant velocity
z1$IV <- aggregate(df$iv, list(df$X), FUN = mean, na.rm=TRUE)[,2] #Mean velocity by larvae
mean_t1_17[[i]]<-z1
}
mean_t1_17<-mean_t1_17[c(26:39)]
names(mean_t1_17)<-list_of_files[26:39] #rownames
# Tidying columns Temp, aquarium and larva with rownames
mean_t1_17<- list.rbind(mean_t1_17)
mean_t1_17$larva<-c(1,1,1,1,2,2,2,2,3,3,3,4,4,4,5,5,5,2,2,3,3,3,3,3,3,3,3,3,3,4,4,5,5,5,5,1,1,1,1,1,1,1,2,2,2,2,2,2,3,3,3,3,3,4,4,4,5,5,5,5,5,5)
mean_t1_17$aquarium<-c("A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "A", "B","B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "B", "C","C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C", "C","C", "C", "C", "C", "C", "C", "C", "C")
mean_t1_17$temp<-rep(17)
mean_t1_17<-mean_t1_17[, c(6,5,4, 1, 2, 3)]
colnames(mean_t1_17)<-c("Temp","aquarium","Larva","Trajectory","Time","IV")
data<-rbind(mean_t1_12,mean_t1_15,mean_t1_17)#Tidying data
data1<-data%>%group_by(Temp,aquarium,Larva) %>%
dplyr::summarise(Ist_Vel = mean(IV),
Tot_Time= sum(Time), Max_Time=max(Time))%>%rowid_to_column(var='ID')
data1$Temp<-as.factor(data1$Temp)#Median of instant velocity by temp
medIV <- ddply(data1, "Temp", summarise, grp.med=median(Ist_Vel), Q1=quantile(Ist_Vel, c(0.25), type = 6),Q3=quantile(Ist_Vel, c(0.75), type = 6), IQR=Q3-Q1)
print(medIV)## Temp grp.med Q1 Q3 IQR
## 1 12 5.138207 3.964596 8.341095 4.376499
## 2 15 1.769275 1.411203 2.419496 1.008293
## 3 17 1.568503 1.257316 2.711036 1.453720#Median of max time swimming by temp
medMax <- ddply(data1, "Temp", summarise, grp.med=median(Max_Time),
Q1=quantile(Max_Time, c(0.25), type = 6),Q3=quantile(Max_Time, c(0.75), type = 6), IQR=Q3-Q1)
print(medMax)## Temp grp.med Q1 Q3 IQR
## 1 12 10.25 5.7675 18.590 12.8225
## 2 15 18.80 9.0000 28.400 19.4000
## 3 17 19.25 11.7000 27.125 15.42502. Kernel plots
#Non-parametric Kernel Densities plot for instant velocity (IV)
dplot_IV<-ggplot(data1, aes(Ist_Vel, color=Temp))+
geom_density(kernel = "gaussian")+
geom_point(aes(x=5.13, y=0),colour="blue",shape=17, size=6)+
geom_point(aes(x=1.76, y=0),colour="orange",shape=17, size=6)+
geom_point(aes(x=1.56, y=0),colour="red",shape=17, size=6)+
theme_classic()+
labs(colour="Temperature", x="Instant Velocity (mm/s)", y="Kernel density")+
scale_x_continuous(limits = c(0, NA), expand = expansion(mult = c(0, 0.1))) +
scale_y_continuous(limits = c(0, NA),expand = expansion(mult = c(0, 0.1))) +
theme(axis.title = element_text(size = 20),axis.text = element_text(size = 20)) +
scale_colour_manual(values =c("blue","darkorange", "red"))
dplot_IV
Kernel density plot of instant velocity (IV) with their median (triangle) by temperature
#Non-parametric Kernel Densities plot for max time swimming (MT)
dplot_Max<-ggplot(data1, aes(Max_Time, color=Temp))+
geom_density(kernel="gaussian")+
geom_point(aes(x=10.25, y=0),color="blue",shape=17, size=6)+
geom_point(aes(x=18.80, y=0),color="orange",shape=17, size=6)+
geom_point(aes(x=19.25, y=0),color="red",shape=17, size=6)+
theme_classic()+
labs(colour="Temperature", x="Max Time Swimming (sec)", y="Kernel density")+
scale_x_continuous(limits = c(0, NA), expand = expansion(mult = c(0, 0.1))) +
scale_y_continuous(limits = c(0, NA),expand = expansion(mult = c(0, 0.1))) +
theme(axis.title = element_text(size = 20),axis.text = element_text(size = 20)) +
scale_color_manual(values =c("blue","darkorange", "red"))
dplot_Max
Kernel density plot of max time swimming (MT) with their median (triangle) by temperature
3. Linear mixed effects model with log transformed data
#Checking for normality
shapiro.test(log(data1$Ist_Vel))##
## Shapiro-Wilk normality test
##
## data: log(data1$Ist_Vel)
## W = 0.96203, p-value = 0.2082hist(log(data1$Ist_Vel))
shapiro.test(log(data1$Max_Time))##
## Shapiro-Wilk normality test
##
## data: log(data1$Max_Time)
## W = 0.96391, p-value = 0.2402hist(log(data1$Max_Time))
#Checking for heteroscedasticity
leveneTest(log(data1$Ist_Vel), data1$Temp)## Levene's Test for Homogeneity of Variance (center = median)
## Df F value Pr(>F)
## group 2 0.2995 0.743
## 36leveneTest(log(data1$Max_Time),data1$Temp)## Levene's Test for Homogeneity of Variance (center = median)
## Df F value Pr(>F)
## group 2 0.2195 0.804
## 36#Linear mixed effects model with Log transformed data of IV. aquarium is a factor with random effects and larva is nested in aquarium
modIV<-lmer(log(Ist_Vel)~Temp + (aquarium|Larva), data=data1)
summary(modIV)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Ist_Vel) ~ Temp + (aquarium | Larva)
## Data: data1
##
## REML criterion at convergence: 60.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6676 -0.5143 -0.0327 0.5135 1.9865
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## Larva (Intercept) 0.039770 0.19942
## aquariumB 0.004512 0.06717 1.00
## aquariumC 0.002630 0.05128 -1.00 -1.00
## Residual 0.228902 0.47844
## Number of obs: 39, groups: Larva, 5
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.6338 0.1566 10.4830 10.435 7.26e-07 ***
## Temp15 -1.1488 0.1961 32.8937 -5.857 1.49e-06 ***
## Temp17 -1.0795 0.1814 31.4611 -5.952 1.33e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Temp15
## Temp15 -0.550
## Temp17 -0.590 0.459
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')anova(modIV)## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Temp 10.937 5.4685 2 32.411 23.89 4.209e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#model diagnostic
plot(modIV,which = 1)
qqnorm(resid(modIV))
qqline(resid(modIV))
#Linear mixed effects model with Log transformed data of MT aquarium is a factor with random effects and larva is nested in aquarium
modMax<-lmer(log(Max_Time)~Temp + (aquarium|Larva), data=data1)
summary(modMax)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: log(Max_Time) ~ Temp + (aquarium | Larva)
## Data: data1
##
## REML criterion at convergence: 73.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.53931 -0.63900 -0.03598 0.69092 1.80417
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## Larva (Intercept) 0.113135 0.33636
## aquariumB 0.009485 0.09739 -1.00
## aquariumC 0.045579 0.21349 -1.00 1.00
## Residual 0.331919 0.57612
## Number of obs: 39, groups: Larva, 5
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 2.4600 0.1831 14.4326 13.436 1.48e-09 ***
## Temp15 0.4205 0.2355 33.1850 1.785 0.0833 .
## Temp17 0.4106 0.2181 32.1673 1.882 0.0689 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) Temp15
## Temp15 -0.562
## Temp17 -0.603 0.461
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')anova(modMax)## Type III Analysis of Variance Table with Satterthwaite's method
## Sum Sq Mean Sq NumDF DenDF F value Pr(>F)
## Temp 1.5309 0.76543 2 32.809 2.3061 0.1156#model diagnostic
plot(modMax,which = 1)
qqnorm(resid(modMax))
qqline(resid(modMax))
4. Pairwise comparisions and log transformed data plots
#pairwise comparison using estimated means of instant velocity by temp
meansIV<-emmeans(modIV, pairwise ~ Temp, adjust="bonferroni")
meansIV## $emmeans
## Temp emmean SE df lower.CL upper.CL
## 12 1.634 0.198 11.3 1.200 2.068
## 15 0.485 0.215 13.5 0.022 0.948
## 17 0.554 0.195 11.1 0.125 0.984
##
## Degrees-of-freedom method: kenward-roger
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## Temp12 - Temp15 1.1488 0.206 24.8 5.584 <.0001
## Temp12 - Temp17 1.0795 0.185 23.4 5.827 <.0001
## Temp15 - Temp17 -0.0693 0.206 24.7 -0.336 1.0000
##
## Degrees-of-freedom method: kenward-roger
## Results are given on the log (not the response) scale.
## P value adjustment: bonferroni method for 3 tests#pairwise comparison using estimated means of max time swimming by temp
meansMax<-emmeans(modMax, pairwise ~ Temp, adjust="bonferroni")
meansMax## $emmeans
## Temp emmean SE df lower.CL upper.CL
## 12 2.46 0.231 11.8 1.96 2.96
## 15 2.88 0.253 14.2 2.34 3.42
## 17 2.87 0.228 11.8 2.37 3.37
##
## Degrees-of-freedom method: kenward-roger
## Results are given on the log (not the response) scale.
## Confidence level used: 0.95
##
## $contrasts
## contrast estimate SE df t.ratio p.value
## Temp12 - Temp15 -0.42051 0.248 25.0 -1.696 0.3067
## Temp12 - Temp17 -0.41056 0.223 23.5 -1.840 0.2354
## Temp15 - Temp17 0.00995 0.249 24.9 0.040 1.0000
##
## Degrees-of-freedom method: kenward-roger
## Results are given on the log (not the response) scale.
## P value adjustment: bonferroni method for 3 tests#Boxplot of logaritmic IV data
p1<-ggplot(data1, aes(x=Temp, y=log(Ist_Vel), fill=Temp))+geom_boxplot()+
stat_summary(fun.y =mean,geom="point",width=0.75,
size=1,linetype="black") +
geom_jitter(color="black", size=0.6, alpha=0.3)+
theme_classic()+labs(x="Temperature", y="IV")+theme(legend.position="none")+
scale_fill_manual(values =c("blue","darkorange", "red"))+
scale_y_continuous(breaks =c(1,2,0))+ geom_text(x=1.2,y=2.3,label="a",color="red")+
geom_text(x=2.2,y=1.3,label="b",color="red")+
geom_text(x=3.2,y=1.3,label="b",color="red")
p1
Boxplot that represents logarithmic IV distribution of the data by temperature, letters indicate significant difference.
# Boxplot of logaritmic MT data
p2<-ggplot(data1, aes(x=Temp, y=log(Max_Time), fill=Temp))+
geom_boxplot()+ stat_summary(fun.y =mean,geom="point",width=0.75,
size=1,linetype="black") +
geom_jitter(color="black", size=0.6, alpha=0.3)+
theme_classic() + theme(legend.position="none") +
labs(x="Temperature", y="MTS")+
scale_fill_manual(values =c("blue","darkorange", "red"))+
geom_text(x=1.2,y=3,label="a",color="red")+
geom_text(x=2.2,y=3.5,label="a",color="red")+
geom_text(x=3.2,y=3.5,label="a",color="red")
p2
Boxplot that represents logarithmic MT distribution of the data by temperature, letters indicate significant difference.
#kernel density plot with inset boxplot of IV
plot.with.inset.IV <-
ggdraw() +
draw_plot(dplot_IV) +
draw_plot(p1, x = .49, y = 0.65, width = .3, height = .35)
plot.with.inset.IV
Fig 4. Kernel density plot of instant velocity swimming by larvae (IV) with their median (triangle) by temperature, also with an inside boxplot that represents the logarithmic IV by temperature.
#kernel density plot with inset boxplot of MT
plot.with.inset.Max_Time <-
ggdraw() +
draw_plot(dplot_Max) +
draw_plot(p2, x = .49, y = 0.65, width = .3, height = .35)
plot.with.inset.Max_Time
Fig 5. Kernel density plot of max time swimming by larvae (MTS) with their median (triangle) by temperature, also with an inside boxplot that represents the logarithmic MTS by temperature.